Filter OperationsApril 23rd & 24th, 2013

Jeremy Hise, P.E.Hazen and Sawyer P.C.

Chesapeake Section AWWA, Inc.Pretreatment and Filter Maintenance,Configuration, and Optimization Seminar

Overview

• Learning objectives – Filtration Background

– Filter Design Considerations

– Why should operators optimize filtration?

– How to measure the progress of optimization?

– What are the important parameters?

– What tools and techniques are needed?

• Necessary skills to assess filter efficiency• Parameters used for comparison

History of Rapid Gravity Filters

• First used at the turn of 20th century

• Used as “roughing” filters as pre-treatment for slow sand filters e.g. London, England

• Used after sedimentation before floc formation and settling was well understood e.g. Cincinnati and Detroit in the US, and Zurich, Switzerland

1905 Richard Miller WTP, Cincinnati, OH

Filtration – Where is it?

• Purpose - remove particulate material from water

DistributionRapid Mix Floc/SedDual-Media

Filter Storage

DistributionRapid Mix FlocDual-Media

Filter Storage

Conventional Treatment

Direct Filtration

Filtration – What does it look like?

What’s missing???

History of Water FiltrationFiltered Water Turbidity Standards

10.0

5.0

1.00.5 0.3 0.1

0.0

2.0

4.0

6.0

8.0

10.0

12.0

Turb

idity

(N

TU)

IESWTR Performance Standards

• Turbidity Performance Requirements– Combined FE Turbidity must be < 0.3 NTU in 95%

of measurements, and never > 1 NTU

• Individual Filter Requirements– Continuous monitoring required for each filter, and

exceptions reported

Filter Backwash Recycle Rule

• FBRR applies to surface water and GWUDI utilities that:– Recycle thickener supernatant– Recycle spent filter backwash water– Recycle dewatering system liquid streams

• There is no size limitation for plants• Rule requires that these streams be returned to

a location prior to all conventional processes

Filtration PhasesRipening

Clean BedRemoval

Breakthrough

Steady stateFiltration

Headloss

Turb

idity

or P

artic

le C

ount

Filter Run Time

What Else Do Filters Do??

• We ask filters to do MORE than filter….– GAC – taste & odor control– Mn treatment– Biological treatment

• Especially for preozone

Filters and Particle SizeParticle Diameter

10-10 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 m

1 Å 1 nm 10 nm 100 nm 1 m 10 m 100 m 1 mm

Molecules

Colloids

Suspended ParticlesVirus

BacteriaAlgae

Cryptosporidium oocystsGiardia cysts

1 Å 1 nm 10 nm 100 nm 1 m 10 m 100 m 1 mm

(After Stumm, ES&T, Vol. 11, p. 1066, 1977)

Nano-particles

Micro-particles

RO

NF UFMF

Granular Media

10 µm(0.01 mm)

To address size & shape of particles desired to be removed

• Create environment in filter media for removal– Physical, but mostly– CHEMICAL

• Chemical influences – Source water quality– Pretreatment– Softening– Recarbonation– Prefilter chemicals

The Prediction of Filtration Performance• Particle Removal and Filter Effluent

Quality; Head loss increase; A Complex Function of:Site Specific & Uncontrolled– Ionic composition– Temperature– Influent Particle sizes, surface

properties, shape and concentration

– Deposit morphology/porosity– Detachment

Designer Controlled– Filtration Rate– Media: size, depth, material,

surface characteristics, porosity– Coagulants/flocculants

What does this all mean?

• Filter media– Provides pore spaces to collect particles

• Collected particles– Accumulate– Need to be removed– Taken away and dealt with

• Performance depends on– Particle conditions – shape and chemistry– Media / underdrain conditions– Operational techniques

Hydraulic Considerations

• Example Filter Cross Section Layout• Design Velocity Guidelines• Inlet Velocities to Prevent Floc Damage• Inlet Velocities to Prevent Media Scouring• Distribution of Backwash Water• Distribution of Air Scour

Simplified Cross-Section of Filter – Normal Flow Path

Fill

FM

Filtered Water Channel

Filter Inlet Channel and Overflow

Waste Wash-water Outlet Channel

Troughs

Media

Underdrain and Plenum

Filter Gallery

Filtered Water Outlet

0 5 ft Filter to Waste with air break

Air Scour Header

Backwash Header

End Gullet

To prevent media scouring ensure inlet gate is not opposite a trough

Design Velocity Guidelines

0 5 ft

Waste Outlet Gate or Valve ~6 ft/s

Inlet Channel ~3 ft/s excluding gate width

Backwash Header ~6 ft/sKeep flooded to prevent air ingress

Air Scour Header, valve and drop pipe ~80 ft/s

Filtered Water Outlet ~6 ft/s, Filter to Waste 8 ft/s

Filter Outlet Channel ~4 ft/s

Backwash Distributers ~4 ft/s – several needed

>10 ft

Discharge Weir above floor

Waste Channel free discharge to tank –velocity can be high

Inlet Gate or Valve ~2 ft/s

Air Scour 3.0 scfm/sq ft

0 5 ft

Air Header above Top Water Level –cannot flood

Air Scour Distributer Header – design varies to suit underdrain type

Water surface 6 inches above media at start, but increases during low rate backwash - affects blower back pressure

Underdrain to give +/- 5% air distribution

Note: SCFM not ACFM

Backwash 6 to 25 gpm/sq ft

0 5 ft

Media expansion15 to 30%, 20% normal

Allow > 1 ft under trough during backwash

Free discharge from troughs and gullet

Backwash distribution directed to floor – note closed end

Underdrain to give better than +/- 5% distribution during backwash

Multi- and Mono-Media Filters• Dual-Media - Anthracite and Sand (> 3 ft) most common• Multi-Media – Lower layer of garnet or ilmenite (FeTiO3)

usually not beneficial – mixes with sand layer• Dual-Media - Sand with GAC cap for taste and odor

removal not organics removal• Mono-Medium Coarse Sand – Used for tertiary filtration

and some overseas water applications for simplicity• Mono-Medium Coarse Anthracite (5 to 6 ft) – not

recommended for potable water, especially if high rate• Mono-Medium GAC – Can be up to 10 ft deep for high

EBCT for organics removal - 12 inches sand underneath to prevent biomass sloughing into filtrate

Media Recommendations• Use AWWA B100-01 Granular Filter Media as basis for

specifications - Read and apply it!!!• Enforce full QA/QC procedures from suppliers premises

to after backwashing and skimming• Do NOT use bulk tanker delivery – severe media attrition

will result plus possible contamination• Use semi-bulk containers / bags of woven material• Protect bags from weather – sun (UV), rain and freezing• Do not stack bulk bags – they will burst• Avoid hydraulic placement if possible – attrition can be

severe leading to more backwashes to remove fines

Filter Media – Key Issues• Filters are the most important part of water treatment

process• Media must be sourced from experienced vendors (e.g.

Unifilt and F.B. Leopold)• Tight QA/QC must be maintained• Media is heavy (4,000 lb) – check structural floor loads,

including fork lift trucks and hoppers• Anthracite is particularly vulnerable to attrition and

incorrect supply (SG, ES, UC etc.)• Effective backwash procedures are crucial for good

media performance

Filter Backwashing

• Filters get cleaned by– Using correct amount of wash water– At the correct flow rate– For the correct amount of time

• Most common problems– Not following procedures – Inconsistent schedules– Poorly designed filters– Poorly designed support facilities– Backwashing based on run time and not

adjusting for water quality

Constant Rate Control - Recommended

L

Mode of OperationInlet water channel level provides flow set point.Valve position controlled by flow signal.As flow to inlet channel increases, water level increases, allowing more flow through all filters. Flow feedback loop maintains constant flow though individual filter. As headloss builds up, the flow drops, causing valve to open to maintain constant flow rate. All filters operate at same flow rate. Avoid control loop “hunting” as this leads to turbidity breakthrough.

Controlled Declining Rate – Avoid in US

L

Mode of OperationWater level in the filter controls valve position directly.As flow to inlet channel increases, water level increases, allowing more flow through all filters. The flow meter is passive and monitors flow only. As headloss builds up, the level increases in the filter, causing valve to open to maintain the same water level for all filters. Filters operate at higher flows when clean and lower flows as the headloss builds up. Filters are usually backwashed on a regular time schedule.

If weir gate is used then constant rate possible. Weirs split flow equally to all filters.

Underdrains and Backwashing Techniques

• Large variety of underdrain designs – lateral and plenum are two main groups

• Backwashing Techniques– Separate air scour + high rate backwash (good)– Combined air scour + low rate backwash, followed by high rate

backwash (best)– High rate backwash with surface sweeps (adequate)

• Waste backwash water removal: – Troughs in USA, – Weirs in Europe

• Backwashing fluidizes the media

Effect of Water Temperature and Viscosity, on Backwash Rate

00.20.40.60.8

11.21.41.61.8

2

0.0 10.0 20.0 30.0 40.0

Absolute Viscosity of Water (cp) vs. Temperature (deg C)

Nearly 30% more

25% less

Media Selection and FluidizationSandSG 2.6 to 2.70ES 0.45 to 0.55 mmUC <1.4UC < 1.4 (critical)

Sand (Average)SG 2.65 ES 0.50 mmES x UC = 0.70 →Backwash = 16.5 gpm/ft2

Anthracite SG 1.55 → 14.0 gpm/ft2

Anthracite UC 1.7 → 19.2 gpm/ft2

Note 20 C 68 FSusuma Kawamura ISBN 0-471-35093-1European anthracite has low SG of 1.4

AnthraciteSG 1.6 to 1.7 (critical)ES 0.8 to 0.9UC < 1.4 (critical)

Anthracite (Average)SG 1.65 ES 0.85ES x UC = 1.19 →Backwash = 15.8 gpm/ft2

Media Expansion, Backwash and Temperature Effects

50%

40%

30%20%

10%

At low temperature and high backwash the result is too high expansion of 50% → media loss

Backwash Operating

Band

Aim for 20% to 30% expansion for sand/anthracite

Example

Troughs – Stainless Steel

Trough support up and down thrust

restraint -adjustable

Lateral brace to prevent vibration

Trough stiffener to prevent vibration

Trough side stiffener (retro)

Media Loss –Common Causes• Uncontrolled air

– Most common• Poor air water distribution

– Possible, watch for spouts • Too high backwash rate

– Possible, watch for churning• Gas bubbles on media

– Unlikely, but dry-bed causes severe foaming

– Biogrowth leading to bubbles on grains

MediaLoss

MediaLoss

No MediaLoss

No MediaLoss

How Air Causes Media LossHow Air Causes Media Loss

Air Scour andRising Wash Limit

Air Scour andRising Wash Limit

Air Scour Only LimitAir Scour Only Limit

Anthracite Anthracite

Sand Sand

Foam after dry bed

Filter Underdrains - Lateral and Plenum

• Lateral (Orifice Based with and without diffusion caps)– F B Leopold (Shown)– AWI– Johnson– Roberts– Severn Trent

• Plenum Concrete Monolithic Floor with Nozzles– Eimco (Shown)– IDI– Orthos

Lateral vs. Plenum FloorPros• Quicker to install• Easier to retro-fit• Pressure contained in

“pipe” lateral – leaks less likely

• Factory built – tends to be more consistent

Cons• Requires skilled

installation• Deep flume disrupts base

slab, makes deeper excavation

• Distribution efficiency depends on lateral length

• No access to clean underdrain if blocked

• Plastic laterals less successful than SS

34

Be careful applying the “capped” versions of lateral underdrains, especially for bio-filters, due to blockage risks

Filter Aid (Polymer)

• Lowers effluent turbidity• Proper dose

– Reduce ripening time– Stabilize turbidity– Stabilize rate of headloss

• Overdose…– May increases rate of headloss development– Take longer to clean

Why Optimize Filters?• Major barrier against pathogen passage• Maximize production efficiency• Minimize spent filter backwash water

– Duration– Frequency

• Increasing reliance on other WQ goals– T&O control– Mn control– Bacteriological stability

How is Progress Measured?

• DATA / TRENDS• Individual filter turbidity• Filter run times• Amount of filter backwash water required• Production efficiency

– What comes in vs. what goes out• Uniform filter run volume (UFRV)

Filters Provide Flexibility, but….

• More Filters - greater chance that:– Bad filter goes unnoticed – Other processes can control filter ops

• Less Filters – greater chance that:– Other filters stressed when one is out of service– Operations less flexible– Plants may operate only part of day

Operator Perspective of Filter Theory

• Filter is particle storage device – not just particle removal device– During storage phase – gentle handling needed– During removal phase – vigorous handling needed

• Filters often designed as dual-media units – Provides deeper bed filtration – Longer runs

• Good filtration depends on good pretreatment – Remember multiple barriers– Short run times = poor efficiency, lots of spent

backwash

Key to Good Filter Operational Techniques

• Continuous operation• At startup

– Bring filter rate up slowly– Don’t start a dirty filter

• During filter run– Ensure filter applied water is stable– Avoid or minimize hydraulic shock– Monitor headloss, NTU, run time– Use filter aid if appropriate for conditions

• After backwash– Rest the filter before returning to service, or– Filter to waste

Key to Good Filter Backwash Techniques

• Prior to backwash– Record filter run information– Verify backwash program parameters

• During backwash– Choose a temperature dependent high flow wash rate– Avoid washes that are too short or long– Hose down the side walls and pipes/gutters– OBSERVE THE BACKWASH

• Observations at each backwash– Surface or air wash effectiveness– View surface for boils or “hot spots”– Look for uneven wash areas or uneven troughs

42

Backwash Program

• Drain– Make sure level is low enough to maximize

energy and minimize media loss• Surface wash or air wash

– 3 to 4 minutes is usually sufficient• Low rate – initiates expansion• High rate – expands media, temp dependent• Low rate - restratification

Key to Good Filter Maintenance Techniques

• Once per quarter (per season)– Adjust high flow rate for temperature– Check media expansion – make adjustments– Review unit filter run volume data– Check media depth– Review all filter profiles

• Once per year– Core the filter – solids retention– Send media to lab for sieve analysis– Add media if necessary– BUT – know why it’s being lost

Probing Media Depth to the Gravel Layer

What Parameters are Important?• Parameters to examine

– Media depth and percent expansion– L/de ratio– Unit Filter Run Volume (UFRV)– Solids retention of media– Backwash use / turbidity / temperature– Filter profile– Sieve analyses – ES & UC, loss of mass

• What do these parameters tell us?– Filter health– Process modifications– Backwash procedure modifications

Filter Inspection Techniques• Visual observation of filter surface and

components• Probing media• Solids retention analyses• Core sampling• Sieve analyses / media assessments

47

Tools and Techniques for Inspection

• HEALTH & SAFETY REQUIREMENTS• Review AWWA Standard B100• Measurement Tools

– Shovel, level, 3/8 inch steel rod, tape measure• Coring Tool

– 11/2 inch electrical conduit, 5 foot length, baggies• Expansion Tool

– One-inch interval tubes or cups• Laboratory Instruments and Tests

– Turbidimeter, glassware, balance, sample bottles, baggies• COMMUNICATIONS

Filter Inspection Tools

Media Assessment – L/D ratio

• Bed Depth Measurement (Drained Bed)– Know original specs

• Effective size - Uniformity Coefficient - Depth - L/D ratio– Use a 3/8 inch steel rod to poke into media, or dig

into it to measure depth• If filter is dual or mixed bed, note depth of each strata,

and depth of mixed interface– Check to see if troughs are level, then measure

distance from trough to bed - check for mounding– Calculate L/D ratio - should be >1100 for low NTU

production

Example – L/D

• Dual Media - originally 36 inches 1 mm anthracite and 6 inches 0.5 mm sand

• Measurement shows 32 inches of anthracite, 6 inches of mixed layer, and 3 inches of sand = 41 total inches

• Rough L/D calculation (send media for analysis)– (32in X 25.4)/1mm 813– (6in X 25.4)/0.75 mm 203– (3in X 25.4)/0.5mm 152– Therefore L/D = 1168

Core Sampling for Solids Retention

• Solids retention analysis best way to determine backwash effectiveness

• Use core sampling tool and baggies to obtain depth samples– Take samples at 0-2 inches, 2-6, 6-12,

12-18, 18-24, etc., until all bed strata are sampled

– Sample before and after backwash– Wash 50 grams of each sample with 5

successive 100 mL washes of lab water

– Measure turbidity of each X 2- plot on graph as NTU/100 grams media

Inserting Core Sampling Tubes

53

54

Examining Media Core Samples

Lab Setup for Core Samples

• Turbidimeter• Pan balance• Baggies – before

and after• Glassware• Lab water• Weigh boats or

other plastic cups

Guidelines – After Backwash• < 30 NTU

– Bed is too clean - examine wash rate and length - this bed will not ripen quickly

• 30 - 60 NTU – Well cleaned and ripened bed - no need for action

• 60 - 120 NTU – Slightly dirty bed - reschedule retention analysis soon

• > 120 NTU – Dirty bed - evaluate filter wash system and procedures

• > 300 NTU – Mudball problem - rehab bed

Solids Retention

• Measures the effectiveness of backwash

• Can show too little or too much backwash

• Change in historical solids retention is cause for concern

• Graph results for database

0

20

40

60

80

100

120

140

160

0-2 in 2-6 in 6-12 in 12-18 in 18-24 in 24-30 in

NT

U/1

00 g

ram

s of

med

ia

BeforeAfter

Spent Backwash Turbidity Analyses

• Too little / too much washing is a common problem• After the first coring, and before the bed expansion

measurement and second coring, the washwater turbidity should be measured for duration of wash

• Sample at 1 minute intervals and analyze• Graph results as NTU vs. time• Record all data

– Volume of backwash, rates, – Ramping intervals, operator habits

Washwater Turbidity Plot

• Turbidity vs. Time• Helps prevent

Excessive washing– Wastes washwater– Strips ripening

• AWWA goal of 10 NTU• This filter washed too

long0

50

100

150

200

250

300

1 min 2 min 3 min 4 min 5 min 6 min 7 min 8 min 9 min

NTU of Wash AWWA Standard

Operators Sampling Backwash Water

Bed Expansion Measurement with Expansion Tool

• Check high flow wash rate (seasonally adjusted)

• Desire 20 – 30% expansion• Position and tie down the

expansion tool so that it rests on top of the bed

• Wash bed under normal conditions and observe amount of expansion

Use of Expansion Tool

Backwash Rate Temp Correction

• Bed Expansion Measurement Rate Requirement– Temp (Deg C) Multiply 25 Deg value by– 30 1.09– 25 1.00– 20 0.91– 15 0.83– 10 0.75– 5 0.68

Example – Bed Expansion

• Bed Expansion Measurement with Expansion Tool (Example for 30 inch bed)– dual media bed - need for a ramping rate, and a final rate– initial ramp might be 5 to 10 percent, or about 2-3 inches– observe expansion tool and adjust

• Bed depth measured at 30 inches• Bed expansion tool captured 9 inches• Bed Expansion Measurement calculations

– 9 inches divided by 30 inches = 30% approx

Calculation of ES & UC from sieve analysis

• Sieve pans used for media size analysis

• ES = D10– 90% larger, 10% smaller

• UC = D60 / D10 • Example sieve analysis

for anthracite– ES= 1.2– UC = 1.2

• 1.2 better than 1.4

Unit Filter Run Volume• UFRV - amount of water that is filtered during

the filter run time – should be determined for every filter run– Goal - UFRV of 5,000 gallons per square foot per run

• Same at low rate or high rate

– Excessive UFRVs are risky– Change in historical UFRV cause for concern– Example:

• 600,000 gals per run / 120 square feet = 5,000 UFRV

Sometimes, you can just tell….

Thank you!Questions?